arXiv subjects

Feihong He

Publications and source records attributed to Feihong He.

15 recordsLinked to original sources

arXiv2026

The Sinking Statistics of Dark Matter Subhalos Across Hierarchical Levels

We investigate the mergers among subhalos in a $\Lambda$CDM simulation, focusing on two fundamental aspects overlooked by previous studies: 1) how to identify mergers robustly; 2) the statistics of mergers across the subhalo hierarchy. To this end, we make use of the HBT+ subhalo finder that tracks subhalo evolution across hierarchy levels, identifying the coalescence of subhalo cores in phase space as a "sinking" event. This coalescence marks a distinct stalled phase in orbital decay, providing a physically motivated and natural definition of a resolved merger. Moreover, the phase transition occurs over a very short timescale, making the statistics robust to numerical resolutions. Our main findings are as follows. (1) More than 90% of sinking events occur between adjacent subhalo levels, while cross-level pathways arise from tidal stripping, group accretion, and numerical constraints. (2) Resolved sinking events are predominantly major mergers (mass ratios > 1/10), whereas the contribution from minor mergers decreases with the dynamical age of the host halo. (3) Although deep-level subhalos typically have low mass ratios relative to the host halo, their mass ratios relative to their direct parents are substantially larger, significantly enhancing their sinking probability. Consequently, the satellite-satellite sinking rate can rival or even exceed the central-satellite sinking rate at lower mass thresholds. (4) Satellite-satellite sinking events are spatially biased toward the outer regions of the host halo, suggesting that the central tidal field suppresses orbital decay between satellite systems.

astro-ph.GA↗

arXiv2026

The tidal evolution of satellite galaxies in cosmological simulations: insights from COLIBRE

We investigate the co-evolution of the stellar and dark matter mass of satellite galaxies using the COLIBRE cosmological hydrodynamical simulations with subhaloes resolved by the history-based HBT-HERONS subhalo finder. We identify a universal tidal track connecting stellar mass loss to subhalo mass loss characterized by two distinct phases, which can be well described by the two-parameter model. The initial phase consists primarily of dark matter stripping, whereas stellar stripping becomes significant only after the subhalo bound mass fraction drops below a critical value ($\sim 0.057$). We find a bimodal mass loss rate distribution of subhaloes. In satellites with modest mass loss rates, the stellar mass is largely frozen. By contrast, the galaxy quickly becomes unresolved, along with the dark matter component for the extreme-mass-loss population, naturally explaining the lack of ``orphan'' galaxies in previous hydrodynamical simulations. Our model also predicts the formation condition for dark-matter-deficient galaxies (DMDGs), whose abundance peaks at $m_{*}\sim 10^{9.5}\,\rm{M}_{\odot}$. The abundance of DMDGs can be very sensitive to numerical effects, with COLIBRE resolving a much larger DMDG population than previous hydrodynamical simulations. We also estimate the influence of artificial disruption on the satellite stellar mass function, which can amount to 20 (50) per cent at $m_* \sim 10^{9} (10^{8}) \, \rm M_\odot$, given a baryonic mass resolution of $\sim 10^{6}\,\rm{M}_{\odot}$.

astro-ph.GA↗

arXiv2025

The shape-velocity alignment of satellites forged by tidal locking and dynamical friction

Utilizing the TNG50 simulation, we study two types of alignments for satellites/subhalos: 1) the alignment of their major axes with the galactocentric radial directions (radial alignment), and 2) with the motion directions (orbital alignment). We find that radial alignment is substantially stronger than orbital alignment, with both signals being consistently stronger for subhalos than for satellites. Interestingly, inward- and outward-moving satellites/subhalos show contrasting orbital alignment behaviors, which can be understood in terms of their radial alignment, orbit decay due to dynamical friction and the effect of tidal stripping. The orbital alignment is stronger in more massive halos. In the end, we explore the orbital alignment measured by a mock observer, and find that the observed alignment for MW satellites is due to projection effects, as the major axes of satellites lie within their orbital planes, approximately coplanar with the observer.

astro-ph.GA↗

arXiv2025

A semi-analytical mock galaxy catalog for the CSST extragalactic surveys from the Jiutian simulations

We introduce a mock galaxy catalog built for the CSST extragalactic surveys using the primary runs of the Jiutian $N$-body simulation suites. The catalogs are built by coupling the GAlaxy Evolution and Assembly (GAEA) semi-analytical model of galaxy formation with merger trees extracted from the simulations using the Hierarchical Bound-Tracing (HBT+) algorithm. The spectral energy distributions (SEDs) and broadband magnitudes are computed using the neural-network-based stellar population synthesizer StarDuster, which is trained on radiative transfer simulations to account for detailed galaxy geometry in modeling dust obscuration. Galaxy light-cones up to $z=5$ are subsequently generated with the BLiC light-cone builder which interpolates the properties of galaxies over time using an optimized interpolation scheme. The resulting catalogs exhibit good convergence in many statistical properties of the galaxy population produced from two different resolution simulations. The catalogs reproduce a number of observed galaxy properties across a range of galaxy mass and redshift, including the stellar mass functions, the luminosity function, gas mass fraction, galaxy size-mass relation and galaxy clustering. We also present the photometric and redshift distributions of galaxies expected to be observed in the CSST surveys.

astro-ph.GA↗

arXiv2025

ExAct: A Video-Language Benchmark for Expert Action Analysis

We present ExAct, a new video-language benchmark for expert-level understanding of skilled physical human activities. Our new benchmark contains 3521 expert-curated video question-answer pairs spanning 11 physical activities in 6 domains: Sports, Bike Repair, Cooking, Health, Music, and Dance. ExAct requires the correct answer to be selected from five carefully designed candidate options, thus necessitating a nuanced, fine-grained, expert-level understanding of physical human skills. Evaluating the recent state-of-the-art VLMs on ExAct reveals a substantial performance gap relative to human expert performance. Specifically, the best-performing GPT-4o model achieves only 44.70% accuracy, well below the 82.02% attained by trained human specialists/experts. We believe that ExAct will be beneficial for developing and evaluating VLMs capable of precise understanding of human skills in various physical and procedural domains. Dataset and code are available at https://texaser.github.io/exact_project_page/

cs.CV↗

arXiv2025

The Jiutian simulations for the CSST extra-galactic surveys

We provide an overview of the Jiutian simulations, a hybrid simulation suite for the China Space Survey Telescope (CSST) extragalactic surveys. It consists of four complementary modules: the primary runs with high resolutions with the fiducial concordance cosmology, the emulator runs exploring the parameter uncertainties around the fiducial cosmology, the reconstruction runs intended for recovering the observed Universe position by position, and the extension runs employing extended cosmologies beyond the standard model. For the primary runs, two independent pipelines are adopted to construct subhaloes and merger trees. On top of them, four sets of mock galaxy light-cone catalogs are produced from semi-analytical models and subhalo abundance matching, providing a variety of observational properties including galaxy SED, emission lines, lensing distortions, and mock images. The 129 emulator runs are used to train the CSST emulator, achieving one percent accuracy in predicting the matter power spectrum over $k\leq 10h{\rm Mpc}^{-1}$ and $z\leq 2$. The reconstruction runs employ a number of subgrid baryonic models to predict the evolution and galaxy population resembling certain regions in the real Universe with constrained initial conditions, enabling controlled investigation of galaxy formation on top of structure formation. The extension runs cover models with warm dark matter, $f(R)$ gravity, interacting dark energy, and nonzero neutrino masses, revealing differences in the cosmic structure under alternative cosmological models. We introduce the specifications for each run, the data products derived from them, the corresponding pipeline developments, and present some main tests. Using the primary runs, we also show that the subhalo peak mass functions of different levels are approximately universal. These simulations form a comprehensive and open library for CSST surveys and beyond.

astro-ph.CO↗

arXiv2025

Self-similar decomposition of the hierarchical merger tree of dark matter halos

In the $\Lambda$CDM universe, structure formation is generally not a self-similar process, while some self-similarity remains in certain statistics which can greatly simplify our description and understanding of the cosmic structures. In this work, we show that the merger tree of dark matter halos is approximately self-similar by investigating the universality of the subhalo peak mass function (PMF) describing the mass distribution of progenitor halos. Using a set of cosmological simulations and identifying subhalos of different merger levels with HBT+, we verify that the level-1 subhalo PMF is close to universal across halo mass, redshift, and cosmology. This approximate self-similarity allows us to analytically derive the subhalo PMF for subhalos accreted at any level (i.e., for sub-sub...halos) through self-convolutions of the level-1 PMF, and the resulting model shows good agreement with simulation measurements. We further derive a number of analytical properties on the hierarchical origin of subhalos. We show that higher-level subhalos dominate at progressively lower peak mass in the PMF and are more likely to originate from major mergers than lower-level ones. At a given merger mass ratio, the subhalo accretion rates at each level track the growth rate of the host halo. At a fixed final mass ratio, however, subhalos of higher-level, higher-mass-ratio and in more massive haloes tend to be accreted more recently. Matching subhalo peak mass to galaxy mass, these results have direct implications on the hierarchical origin of satellite galaxies.

astro-ph.CO↗

arXiv2024

Why artificial disruption is not a concern for current cosmological simulations

Recent studies suggest that cold dark matter subhalos are hard to disrupt and almost all cases of subhalo disruption observed in numerical simulations are due to numerical effects. However, these findings primarily relied on idealized numerical experiments, which do not fully capture the realistic conditions of subhalo evolution within a hierarchical cosmological context. Based on the Aquarius simulations, we identify clear segregation in the population of surviving and disrupted subhalos, which corresponds to two distinct acquisition channels of subhalos. We find that all of the first-order subhalos accreted after redshift 2 survive to the present time without suffering from artificial disruption. On the other hand, most of the disrupted subhalos are sub-subhalos accreted at high redshift. Unlike the first-order subhalos, sub-subhalos experience pre-processing and many of them are accreted through major mergers at high redshift, resulting in very high mass loss rates. We confirm these high mass loss rates are physical through both numerical experiments and semi-analytical modeling, thus supporting a physical origin for their rapid disappearance in the simulation. Even though we cannot verify whether these subhalos have fully disrupted or not, their extreme mass loss rates dictate that they can at most contribute a negligible fraction to the very low mass end of the subhalo mass function. We thus conclude that current state-of-the-art cosmological simulations have reliably resolved the subhalo population.

astro-ph.GA↗

arXiv2024

FineDiffusion: Scaling up Diffusion Models for Fine-grained Image Generation with 10,000 Classes

The class-conditional image generation based on diffusion models is renowned for generating high-quality and diverse images. However, most prior efforts focus on generating images for general categories, e.g., 1000 classes in ImageNet-1k. A more challenging task, large-scale fine-grained image generation, remains the boundary to explore. In this work, we present a parameter-efficient strategy, called FineDiffusion, to fine-tune large pre-trained diffusion models scaling to large-scale fine-grained image generation with 10,000 categories. FineDiffusion significantly accelerates training and reduces storage overhead by only fine-tuning tiered class embedder, bias terms, and normalization layers' parameters. To further improve the image generation quality of fine-grained categories, we propose a novel sampling method for fine-grained image generation, which utilizes superclass-conditioned guidance, specifically tailored for fine-grained categories, to replace the conventional classifier-free guidance sampling. Compared to full fine-tuning, FineDiffusion achieves a remarkable 1.56x training speed-up and requires storing merely 1.77% of the total model parameters, while achieving state-of-the-art FID of 9.776 on image generation of 10,000 classes. Extensive qualitative and quantitative experiments demonstrate the superiority of our method compared to other parameter-efficient fine-tuning methods. The code and more generated results are available at our project website: https://finediffusion.github.io/.

cs.CV↗

arXiv2024

FreeStyle: Free Lunch for Text-guided Style Transfer using Diffusion Models

The rapid development of generative diffusion models has significantly advanced the field of style transfer. However, most current style transfer methods based on diffusion models typically involve a slow iterative optimization process, e.g., model fine-tuning and textual inversion of style concept. In this paper, we introduce FreeStyle, an innovative style transfer method built upon a pre-trained large diffusion model, requiring no further optimization. Besides, our method enables style transfer only through a text description of the desired style, eliminating the necessity of style images. Specifically, we propose a dual-stream encoder and single-stream decoder architecture, replacing the conventional U-Net in diffusion models. In the dual-stream encoder, two distinct branches take the content image and style text prompt as inputs, achieving content and style decoupling. In the decoder, we further modulate features from the dual streams based on a given content image and the corresponding style text prompt for precise style transfer. Our experimental results demonstrate high-quality synthesis and fidelity of our method across various content images and style text prompts. Compared with state-of-the-art methods that require training, our FreeStyle approach notably reduces the computational burden by thousands of iterations, while achieving comparable or superior performance across multiple evaluation metrics including CLIP Aesthetic Score, CLIP Score, and Preference. We have released the code at: https://github.com/FreeStyleFreeLunch/FreeStyle.

cs.CV↗

arXiv2024

Mementos: A Comprehensive Benchmark for Multimodal Large Language Model Reasoning over Image Sequences

Multimodal Large Language Models (MLLMs) have demonstrated proficiency in handling a variety of visual-language tasks. However, current MLLM benchmarks are predominantly designed to evaluate reasoning based on static information about a single image, and the ability of modern MLLMs to extrapolate from image sequences, which is essential for understanding our ever-changing world, has been less investigated. To address this challenge, this paper introduces Mementos, a new benchmark designed to assess MLLMs' sequential image reasoning abilities. Mementos features 4,761 diverse image sequences with varying lengths. We also employ a GPT-4 assisted method to evaluate MLLM reasoning performance. Through a careful evaluation of nine recent MLLMs on Mementos, including GPT-4V and Gemini, we find that they struggle to accurately describe dynamic information about given image sequences, often leading to hallucinations/misrepresentations of objects and their corresponding behaviors. Our quantitative analysis and case studies identify three key factors impacting MLLMs' sequential image reasoning: the correlation between object and behavioral hallucinations, the influence of cooccurring behaviors, and the compounding impact of behavioral hallucinations. Our dataset is available at https://github.com/umd-huang-lab/Mementos.

cs.CV↗

arXiv2023

PrototypeFormer: Learning to Explore Prototype Relationships for Few-shot Image Classification

Few-shot image classification has received considerable attention for overcoming the challenge of limited classification performance with limited samples in novel classes. Most existing works employ sophisticated learning strategies and feature learning modules to alleviate this challenge. In this paper, we propose a novel method called PrototypeFormer, exploring the relationships among category prototypes in the few-shot scenario. Specifically, we utilize a transformer architecture to build a prototype extraction module, aiming to extract class representations that are more discriminative for few-shot classification. Besides, during the model training process, we propose a contrastive learning-based optimization approach to optimize prototype features in few-shot learning scenarios. Despite its simplicity, our method performs remarkably well, with no bells and whistles. We have experimented with our approach on several popular few-shot image classification benchmark datasets, which shows that our method outperforms all current state-of-the-art methods. In particular, our method achieves 97.07\% and 90.88\% on 5-way 5-shot and 5-way 1-shot tasks of miniImageNet, which surpasses the state-of-the-art results with accuracy of 0.57\% and 6.84\%, respectively. The code will be released later.

cs.CV↗

arXiv2023

Cartoondiff: Training-free Cartoon Image Generation with Diffusion Transformer Models

Image cartoonization has attracted significant interest in the field of image generation. However, most of the existing image cartoonization techniques require re-training models using images of cartoon style. In this paper, we present CartoonDiff, a novel training-free sampling approach which generates image cartoonization using diffusion transformer models. Specifically, we decompose the reverse process of diffusion models into the semantic generation phase and the detail generation phase. Furthermore, we implement the image cartoonization process by normalizing high-frequency signal of the noisy image in specific denoising steps. CartoonDiff doesn't require any additional reference images, complex model designs, or the tedious adjustment of multiple parameters. Extensive experimental results show the powerful ability of our CartoonDiff. The project page is available at: https://cartoondiff.github.io/

cs.CV↗

arXiv2023

Extending the unified subhalo model to warm dark matter

Using a set of high-resolution N-body simulations, we extend the unified distribution model of cold dark matter (CDM) subhaloes to the warm dark matter(WDM) case. The same model framework combining the unevolved mass function, unevolved radial distribution, and tidal stripping can predict the mass function and spatial distribution of subhaloes in both CDM and WDM simulations. The dependence of the model on the DM particle property is universally parameterized through the half-mode mass of the initial power spectrum. Compared with the CDM model, the WDM model differs most notably in two aspects. 1) In contrast to the power-law form in CDM, the unevolved subhalo mass function for WDM is scale-dependent at the low mass end due to the cut-off in the initial power spectrum. 2) WDM subhaloes are more vulnerable to tidal stripping and disruption due to their lower concentrations at accretion time. Their survival rate is also found to depend on the infall mass. Accounting for these differences, the model predicts a final WDM subhalo mass function that is also proportional to the unevolved subhalo mass function. The radial distribution of WDM subhaloes is predicted to be mass-dependent. For low mass subhaloes, the radial distribution is flatter in the inner halo and steeper in the outer halo compared to the CDM counterpart, due to the scale-dependent unevolved mass function and the enhanced tidal stripping. The code for sampling subhaloes according to our generalized model is available at https://github.com/fhtouma/subgen2 .

astro-ph.CO↗

arXiv2022

Is the core-cusp problem a matter of perspective: Jeans Anisotropic Modeling against numerical simulations

Mock member stars for 28 dwarf galaxies are constructed from the cosmological Auriga simulation, which reflect the dynamical status of realistic stellar tracers. The axis-symmetric Jeans Anisotropic Multi-Gaussian Expansion (JAM) modeling is applied to 6,000 star particles for each system, to recover the underlying matter distribution. The stellar or dark matter component individually is poorly recovered, but the total profile is constrained more reasonably. The mass within the half-mass radius of tracers is recovered the tightest, and the mass between 200 and 300 pc, $M(200-300\mathrm{pc})$, is constrained ensemble unbiasedly, with a scatter of 0.167 dex. If using 2,000 particles and only line-of-sight velocities with typical errors, the scatter in $M(200-300\mathrm{pc})$ is increased by $\sim$50%. Quiescent Sagittarius dSph-like systems and star-forming systems with strong outflows show distinct features, with $M(200-300\mathrm{pc})$ mostly under-estimated for the former, and likely over-estimated for the latter. The biases correlate with the dynamical status, which is a result of contraction motions due to tidal effects in quiescent systems or galactic winds in star-forming systems, driving them out of equilibrium. After including Gaia DR3 proper motion errors, we find proper motions can be as useful as line-of-sight velocities for nearby systems at $<\sim$60 kpc. By extrapolating the actual density profiles and the dynamical constraints down to scales below the resolution, we find the mass within 150 pc can be constrained ensemble unbiasedly, with a scatter of $\sim$0.255 dex. In the end, we show that the contraction of member stars in nearby systems is detectable based on Gaia DR3 proper motion errors.

astro-ph.GA↗